Cloud AI and what it can do.
PENTER Cloud AI and remote command execution are separate features. Here is what you can use during the public beta, how your account applies, and what to do if a request is unavailable.
Cloud AI isn't the same as remote scanning
| Feature | What it means |
|---|---|
| Cloud AI chat | Your account can access supported hosted AI models and agent reasoning when the service and plan allow it. |
| Desktop Browser & HTTP tools | Scoped, local Desktop tools inspect an authorized target under PENTER execution controls. |
| Cloud command execution | Remote security-tool commands are a separate capability. This runner is not enabled for the public beta. |
Which execution features are currently supported?
Use PENTER Desktop's scoped Browser, HTTP and Request Lab for authorized target-facing work. Native shell, WSL/Kali command execution, MCP execution and networked managed Docker are disabled in packaged public-beta builds while destination-scoped controls are evaluated.
- Set your active authorized engagement target before using Browser or HTTP tools.
- Review approvals and inspect actual tool results rather than trusting an AI summary of an unexecuted step.
- Keep managed Docker containers offline with network=none if you choose to use that optional feature.
- Do not attempt to bypass a blocked feature or scope restriction.
Models, plans and usage
Your available hosted models and usage depend on your PENTER account and any active plan allowances. Visit the dashboard to check current model access and recent usage. Cloud AI messages may be sent to the configured provider; avoid including secrets or unauthorized customer data in prompts.
If your Cloud request fails
More step-by-step help is available in the Troubleshooting guide.
Ready to get started?
Follow the setup guide, install the current approved Desktop release, and connect your account.